Collaborative Graph Contrastive Learning without Handcrafted Graph Data Augmentations
A novel collaborative graph contrastive learning framework (CGCL) that generates contrastive views from multiple graph encoders without relying on handcrafted data augmentations.
